Research Of A Proxy Caching Model For Streaming Media Based On Segmentation Strategy

With the rapid increase in the bandwidth available to end users and Web sites hosting user generated content, there appears to be more streaming media content on the Web than ever before. Improve the quality of delivering streaming media content and increase the performance of Web service are become a hot research topic all over the world in recent years.Proxy caching is one of the key techniques to accelerate streaming media delivery. In order to reduce high network resource consumption, streaming media systems usually utilize proxy server and effective cache algorithms. Different from conventional Web objects, media objects have large volume and high bandwidth demand, so full-object caching isn't feasible. Generally, we use segmented cache method to cache streaming media. Meanwhile, exponential segmentation and Adaptive & Lazy segmentation are typical segmented-cache strategies. They can reduce backbone Internet traffic and server workload effectively.The performance metrics of streaming media cache system act as an important role on measuring the segmentation and replacement. This paper surveys a number of performance metrics for streaming media and points out their disadvantages, then according to computing much more complex of caching media objects, this article puts forward the concept of object removal rate and applies it to the effect of CPU and disk usage when adopting different replacement algorithm. The concept can provide suggestion and reference for designing cache system.This paper develops the caching replacement utility function of Adaptive & Lazy strategy. In order to make total available data volume for cache, the function takes every segment's popularity into consideration and makes the caching data proportional with the popularity. Simulation proves that, by employing the developed algorithm, the byte hit rate is improved and delayed response rate is decreased. The algorithm shows good performance in the log of high access frequency.